Bitcoin Miners Shift to AI Infrastructure Amid Declining Margins and New Revenue Streams
Autor: Mining Provider Editorial Staff
Veröffentlicht:
Kategorie: News
Zusammenfassung: Bitcoin miners are increasingly converting their infrastructure to AI data centers, driven by the need for stable revenue amid declining mining margins and securing multi-year contracts with AI firms. This shift allows them to leverage existing assets while balancing risks associated with hardware depreciation and operational challenges.
Bitcoin Miners Transition to AI Infrastructure – Initial Effects Observed
Bitcoin mining companies are increasingly converting a significant portion of their physical infrastructure to AI data center operations. This transition, once considered experimental, has become a recognized revenue source for several publicly traded miners by 2026.
Miners are making this shift because they possess large-scale energy infrastructure, physical data center space, and cooling systems, which are essential for AI companies to operate GPU clusters. The costs associated with retrofitting existing mining facilities are considerably lower than building entirely new data centers from scratch. Additionally, mining margins have shrunk due to increased difficulty following the halving in April 2024.
“AI hosting contracts offer more predictable revenues than BTC block rewards, which fluctuate with both price and difficulty.”
Several miners have secured multi-year hosting agreements with AI and machine learning companies, ensuring dollar-denominated revenues that are not affected by BTC price volatility.
Key Insights: The shift to AI infrastructure is driven by the need for stable revenue sources amidst declining mining margins. Miners are leveraging existing assets to meet the growing demand for AI computing power.
Leading Companies in the Transition
Hut 8 is one of the most prominent examples of this transition, with its stock rising by 112% in 2026, trading near $97 according to MarketBeat data from May 20. The company has clearly articulated its AI hosting ambitions in public statements and investor communications.
Core Scientific signed a significant AI data center contract with CoreWeave in 2024, setting a benchmark for the entire sector. Other companies like CleanSpark, Cipher Mining, and Riot Platforms have also disclosed varying stages of evaluation or active deployment of AI hosting in recent quarters. A common factor among these miners is that those with newer, higher-density power connections are better positioned to operate modern GPU clusters compared to those with older, less dense sites.
Key Insights: Companies like Hut 8 and Core Scientific are leading the charge in AI hosting, showcasing significant stock performance and strategic partnerships that highlight the industry's shift.
Background of the Shift
The idea of miners diversifying into AI hosting gained widespread attention in late 2023 when Core Scientific's deal with CoreWeave provided a concrete financial model for the concept. Previously, most miners viewed their infrastructure primarily as a single-purpose asset for Bitcoin production.
The 2024 halving, which reduced the block reward from 6.25 BTC to 3.125 BTC, increased the pressure to find alternative revenue sources. Concurrently, the rapid growth in training and inference of AI models led to a GPU shortage, making any owner of spare energy and cooling capacity a potential partner for AI companies.
Key Insights: The transition to AI hosting is a strategic response to the economic pressures of Bitcoin mining, driven by the need for alternative revenue streams and the growing demand for AI capabilities.
The Economics of Dual-Use Infrastructure
A mining facility that mines BTC generates its revenue in BTC, which is then converted to USD at market price. In contrast, a facility operating GPU clusters for an AI client earns a fixed dollar rate per GPU hour under a hosting contract. This dollar-based model is less volatile but can also be less lucrative during BTC bull markets. Miners typically do not abandon Bitcoin mining entirely; instead, they allocate additional capacity for AI rather than further ASICs.
For instance, a miner building a new 100-megawatt facility in 2026 is more likely to split it 50/50 between BTC mining and AI hosting rather than dedicating it entirely to one application. This hedging strategy smooths revenues but also means that miners are no longer seen as pure Bitcoin exposure by equity investors.
Key Insights: The dual-use model allows miners to balance their revenue streams, mitigating risks associated with market volatility while still participating in Bitcoin mining.
Risks and Limitations
The pivot to AI hosting is not without complications. GPU hardware is expensive and can quickly depreciate in value as new chip generations are released. Miners investing in current H100 or H200 clusters may find these assets partially obsolete within 24 months if the next generation from Nvidia is released as scheduled.
While long-term hosting contracts provide revenue certainty, they can also bind miners to hardware that may age poorly. Additionally, there is a skills gap; operating an AI data center requires a different operational profile than running an ASIC mining farm. Several miners have responded by hiring data center engineers from the cloud industry rather than promoting from within their mining teams.
Key Insights: The transition to AI hosting presents significant risks, including hardware depreciation and the need for specialized operational skills, which could impact the viability of this new business line.
Sources: